Skip to content

Latest commit

 

History

History
39 lines (29 loc) · 1.75 KB

TROUBLESHOOTING.md

File metadata and controls

39 lines (29 loc) · 1.75 KB
Raw

Troubleshooting

Hitting one hostname actually tries to serve another

This can happen if you don't use the standard setup-app command

$ tail /usr/local/var/log/nginx/error.log shows something like this

2020/04/03 13:37:13 [error] 51098#0: *5 kevent() reported that connect() failed (61: Connection refused) while connecting to upstream, client: 127.0.0.1, server: promo.thegulocal.com, request: "GET / HTTP/1.1", upstream: "http://[::1]:9500/", host: "mem.thegulocal.com"

you can see the correct host mem.thegulocal.com, but the incorrect server and port number promo.thegulocal.com and 9500.

This is because if nginx can't find a matching server block for host AND protocol, it just picks the first one, which in this case is promo. If you access with https it works fine, but be sure to define the http as well as the https in your nginx config file if you want the redirect to work:

server {
  server_name mem.thegulocal.com;

  location / {
    proxy_pass http://localhost:9100/;
      proxy_set_header Host $http_host;
  }
}

Hash bucket size emerg

Seeing a message such as nginx: [emerg] could not build server_names_hash, you should increase server_names_hash_bucket_size: 64 after starting nginx is a sign that an unusually long server name has been defined.

You'll want to set server_names_hash_bucket_size within the http directive in /usr/local/etc/nginx/nginx.conf to the next power of two. By default, server_names_hash_bucket_size is 32 so 64 or 1024 should be enough:

http {
  server_names_hash_bucket_size  64;
  ...
}

See the docs for more information.